public void SaveToDatabase(string name)
{
PageMap pm;
if (string.IsNullOrEmpty(name))
{
int secondSlash = _baseUrl.IndexOf("/") + 1;
string domain = _baseUrl.Substring(secondSlash + 1);
name = domain.Substring(0, domain.IndexOf("/"));
}
if (!_db.PageMaps.Any(m => m.Name == name && m.BaseURL == _baseUrl))
{
pm = new PageMap();
pm.PageMapID = Guid.NewGuid();
pm.Name = name;
pm.BaseURL = _baseUrl;
_db.PageMaps.InsertOnSubmit(pm);
}
else
{
pm = _db.PageMaps.Single(m => m.Name == name && m.BaseURL == _baseUrl);
_db.Nodes.DeleteAllOnSubmit(_db.Nodes.Where(n => n.PageMapID == pm.PageMapID));
}
int index = -1;
foreach (var node in _nodeMap)
{
index++;
Node newNode = new Node();
newNode.PageMapID = pm.PageMapID;
newNode.NodeIndex = index;
newNode.NodeName = node.Name;
newNode.NodeID = node.Id;
if (node.Attributes.Any(a => a.Name == "class"))
{
newNode.NodeClass = node.Attributes.Single(a => a.Name == "class").Value;
}
_db.Nodes.InsertOnSubmit(newNode);
}
_maxIndex = pm.MaxIndex = index;
_db.SubmitChanges();
}
public static List<string> GetRelevantTextFromDocumentUsingMap(HtmlDocument document, PageMap map)
{
HtmlNode documentNode = document.DocumentNode.SelectSingleNode("//html");
List<HtmlNode> documentNodesOfInterest = new List<HtmlNode>();
List<string> textItems = new List<string>();
foreach (Node mapNode in map.Nodes.OrderBy(n => n.NodeIndex))
{
// If we're at the start of the map, find where to go next
if (mapNode.NodeIndex < map.MaxIndex)
{
// Find out which child nodes have the right name and ID.
documentNodesOfInterest = documentNode.ChildNodes.Where(n => n.Name == mapNode.NodeName && n.Id == mapNode.NodeID).ToList();
// If there's only 1, use that.
if (documentNodesOfInterest.Count == 1)
{
documentNode = documentNodesOfInterest.First();
continue;
}
// If there are more, find out which ones have the right class
else
{
foreach (HtmlNode docNode in documentNodesOfInterest)
{
foreach (HtmlAttribute a in docNode.Attributes.AttributesWithName("class"))
{
if (a.Value == mapNode.NodeClass)
{
documentNode = docNode;
break;
}
}
if (documentNode == docNode)
{
break;
}
}
// If none had the right class, they won't have been put in documentNode
if (!documentNodesOfInterest.Contains(documentNode))
{
// Try to match on the ID of a child node in the map
foreach (HtmlNode docNode in documentNodesOfInterest)
{
Node nextNode = map.Nodes.Where(n => n.NodeIndex > mapNode.NodeIndex && !string.IsNullOrEmpty(n.NodeID)).OrderBy(n => n.NodeIndex).FirstOrDefault();
if (nextNode != null && docNode.InnerHtml.Contains(nextNode.NodeID))
{
documentNode = docNode;
break;
}
}
// If we still haven't found anything, it's usually because the
// loop has hit a listitem. Loop through from here, see what we
// get and don't bother going deeper.
if (!documentNodesOfInterest.Contains(documentNode))
{
foreach (HtmlNode node in documentNodesOfInterest.Where(n => n.Name == mapNode.NodeName && n.Id == mapNode.NodeID))
{
{
textItems.Add(node.InnerText.Trim());
}
}
break;
}
}
}
}
// If we're at the end of the map, start delving into the text
else
{
foreach (HtmlNode node in documentNode.ChildNodes.Where(n => n.Name == mapNode.NodeName && n.Id == mapNode.NodeID))
{
// Without class information, the next bit is useless. Grab any old text.
if (mapNode.NodeClass == null)
{
{
textItems.Add(node.InnerText.Trim());
}
}
else
{
// If we have class information, only pull the matching ones.
foreach (HtmlAttribute a in node.Attributes.AttributesWithName("class"))
{
string className = map.Nodes.Single(n => n.NodeIndex == map.MaxIndex).NodeClass;
if (a.Value == className)
{
textItems.Add(node.InnerText.Trim());
}
}
}
}
}
}
return textItems;
}
partial void UpdatePageMap(PageMap instance);
partial void DeletePageMap(PageMap instance);
partial void InsertPageMap(PageMap instance);
